top of page
  • Katherine Hollingsworth

MedeA - An Atomistic Simulation Environment for ICME

Monday, May 11, 2015: 4:00 PM

Room 201B (Long Beach Convention and Entertainment Center) Dr. Clive Freeman , Materials Design, Inc., Angel Fire, NM Dr. Erich Wimmer , Materials Design, SARL, Montrouge, France Dr. Paul Saxe , Materials Design, Inc., Angel Fire, NM

At its heart, ICME is based on linking models, data and experiment from different scales to provide an integrated approach to understanding and developing complex materials. At the moment, there are many models that are applicable at various length and time scales, but they are not integrated and an overarching issue concerns the origin of the input parameters required by the models. Over the past fifteen years we have been developing MedeA, an atomistic modeling environment that tackles this problem from the bottom up. The advantage of atomistic modeling is that, at the smallest scale, there is no input other than fundamental physical constants. However, ab initio techniques are limited to simulating no more than a few hundred atoms and for very short times, on the order of nanoseconds. Hence we need to focus on taking the output of these ab initiomodels to larger scales, working upwards to provide the inputs for empirical models, which themselves are in part working downwards to smaller scales.

In this talk, we will cover the properties, such as diffusion, elastic constants, and interface energies that can be directly calculated from model systems ab initio, and also recent work on using ab initio data as input to larger models. This includes automatically generating forcefields (potentials) for molecular dynamics, allowing the simulation of hundreds of thousands of atoms, as well as cluster expansion methods to bring the high accuracy and generality of the ab initio calculations to predicting phase diagrams and local ordering in alloys. These methods in turn provide input for CALPHAD methods, yielding an integrated capability based on the smallest of length and time scales, requiring few empirical parameters, and extending to practical and macroscopic scales.

bottom of page